LookaheadOptimizer-mx
LookaheadOptimizer-mx copied to clipboard
wkcn
Lookahead Optimizer: k steps forward, 1step back for MXNet
Results
2
LookaheadOptimizer-mx issues
Sort by
recently updated
recently updated
newest added
thanks a lot for your excellent implementation! I just got an issue when trying to call the _save_states()_ function of gluon trainer after the training finished. error log: ... return...
bug
For optimizers like sgd+momentum, adam, rmsprop, they may use the historical information of the gradients. Does this implementation maintain / reset / interpolate the momentum in each outer loop?
